Replit, the AI coding platform popular with startup founders, is no longer working with Anthropic, chief executive and co-founder Amjad Masad said on stage at TechCrunch Disrupt, in remarks reported by TechCrunch on Friday.

Masad said the decision came down to cost, partnership and performance. Replit announced a partnership with OpenAI in August, but there had been no public indication until now that the Anthropic relationship had ended. Under the Anthropic partnership, Replit used the lab's Claude models to run its platform, an arrangement credited with helping the company grow its revenue significantly.

Masad said Replit uses an internal benchmark, alongside a public one, to assess model performance, and that OpenAI's models performed better on building vibe-coded applications from scratch. 'I think ChatGPT is better now,' he said.

The comments were disputed by people familiar with the Anthropic partnership, one of whom told TechCrunch the relationship was still in place, and Replit's own site and blog cited Anthropic's Claude models in production as recently as late September. TechCrunch said it had asked Replit for clarification.